The Use of Multi-Dimensional Scaling in Policy Selection

Abstract

The selection of the best from a set of available policies with multi-dimensional consequences is a basic problem in decision analysis. In this paper the technique of multi-dimensional scaling is used to define a preference structure from attitude data collected from the decision maker. This data is based on a comparison of each pair of policies. Using the concept of the ideal solution, the absolute preferences for the available policies can be defined, hence allowing the policies to be ranked from most to least desirable.